DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 05/26/2026 has been entered.
Information Disclosure Statement
The information disclosure statements (“IDS”) filed on 07/10/2026 has been reviewed and the listed reference has been considered.
Status of Claims
Claims 1-20 are cancelled. Claims 21-35 are new and pending.
Response to Arguments
Applicant’s response, which has introduced independent Claims 21 and 30, has altered the scope of the claims of the instant application, which has necessitated the new ground(s) of rejection presented in this office action with respect to the new claims of the instant application. Accordingly, in response to Applicant’s arguments that are merely directed to the new claims, new analyses have been presented below, which make Applicant’s arguments moot.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 21-29 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Specifically, Claim 21 recites “the specified dataset”. There is insufficient antecedent basis for “specified dataset” in the claim. Upon review of the claims, claim 24 provides antecedent basis for the limitation.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 21-29 are rejected under 35 U.S.C. 103 as being unpatentable over Mack (US 2019/0272623 A1) in view of Ramos et al. (US 2019/0244113 A1) and in further view of Pandev et al. (US 2019/0214285 A1).
Regarding claim 21, Mack teaches, A non-transitory computer-readable medium having instructions (Mack, ¶0204: “A computer-readable medium storing instructions”) that, when executed by a computer, cause the computer to perform operations (Mack, ¶0204: “instructions that are executable by a processor to cause a computer to execute operations comprising”) for training a machine learning model (Mack, ¶0197: “train the machine learning model”) to determine a source of error contribution (Mack, ¶0057: “stochastic variations, such edge placement errors caused by roughness or stochastic effects”) to multiple features of a pattern printed on a substrate, (Mack, ¶0216: “predict a variability in a final edge position of a printed pattern due to the stochastic effects”) the operations comprising: obtaining training data having multiple datasets, (Mack, ¶0150: “Inputs to the machine-learning algorithm are called training data sets”) wherein each dataset has error contribution values (Mack, ¶0151: “training data sets include unbiased roughness measurements and/or unbiased PSD measurements or parameters”) representative of an error contribution (Mack, ¶0160: “power spectral density (PSD) dataset representing feature geometry information corresponding to the edge detection measurements of the set of images”) from one of multiple sources to the features, (Mack, ¶0045: “sources of variation that affect patterning fidelity (e.g., exposure dose and focus variations, hotplate temperature non-uniformity, scanner aberrations”). However, Mack does not explicitly teach, and wherein each dataset is associated with an actual classification that identifies a source of the error contribution of the corresponding dataset; and training, based on the training data, a machine learning model to predict a classification of a reference dataset of the datasets such that a cost function that determines a difference between the predicted classification and the actual classification of the reference dataset is reduced, wherein the classification identifies a specified source of the multiple sources as the source of error contribution for the error contribution values in the specified dataset.
In an analogous field of endeavor, Ramos teaches, wherein each dataset is associated with an actual classification that identifies (Ramos, ¶0001: “predictions made by the classifier model (as trained based on a labeled training dataset”; the actual classification is interpreted as a label) a source of the error contribution of the corresponding dataset; (Ramos, ¶0037: “error sources can be grouped”) and training, based on the training data, a machine learning model (Ramos, ¶0006: “training data used to train the classifier model”) to predict a classification of a reference dataset of the datasets (Ramos, ¶0001: “predictions made by the classifier model (as trained based on a labeled training dataset”) such that a cost function that determines a difference between the predicted classification and the actual classification of the reference dataset is reduced, (Ramos, ¶0036: “minimizing a cost function that measures the discrepancy between the (e.g., manually assigned) labels and the predictions”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Mack using the teachings of Ramos to introduce computing a cost function of difference between a predicted class and an actual class. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of evaluating the accuracy of a classification model using the cost function. Therefore, it would have been obvious to combine the analogous arts Mack and Ramos to obtain the above-described limitations in claim 21. However, the combination of Mack and Ramos does not explicitly teach, wherein the classification identifies a specified source of the multiple sources as the source of error contribution for the error contribution values in the specified dataset.
In another analogous field of endeavor, Pandev teaches, wherein the classification identifies a specified source (Pandev, ¶0031: “a first defect class corresponds to a first type of defect that is visible on a first wafer”) of the multiple sources as the source of error contribution for the error contribution values in the specified dataset. (Pandev, ¶0033: “the types of defects for the defect classes have been specified (e.g., are known), the defect classes are labelled with labels that specify the types of defects for the defect classes”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Mack in view of Ramos using the teachings of Pandev to introduce specified known defect classes. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of training the classification model to identify specific sources of defects in the semiconductor images. Therefore, it would have been obvious to combine the analogous arts Mack, Ramos and Pandev to obtain the invention in claim 21.
Regarding claim 22, Mack in view of Ramos and in further view of Pandev teaches, The non-transitory computer readable medium of claim 21, wherein obtaining the training data includes: obtaining local critical dimension uniformity (LCDU) data (Mack, ¶0070: “This feature-to-feature variation is called the local critical dimension uniformity, LCDU, since it represents CD (critical dimension) variation”) associated with the features using different focus and dose level values of an apparatus used for printing the pattern. (Mack, ¶0045: “sources of variation that affect patterning fidelity (e.g., exposure dose and focus variations”).
Regarding claim 23, Mack in view of Ramos and in further view of Pandev teaches, The non-transitory computer readable medium of claim 22, wherein obtaining the training data includes: decomposing LCDU data associated with the features to derive the error contribution values from each of the multiple sources. (Mack, ¶0073: “parameters PSD(0), ξ, and H, enables one to predict the stochastic influence on a line of any length L. It is noted that the LCDU does not depend on the roughness exponent, making H less important than PSD(0) and ξ”).
Regarding claim 24, Mack in view of Ramos and in further view of Pandev teaches, The non-transitory computer readable medium of claim 21, wherein the operations further comprise: receiving a specified dataset having error contribution values representative of an error contribution (Mack, ¶0005: “Each image of the set includes one or more instances of a feature within a respective pattern structure, and each image includes measured linescan information corresponding to the pattern structure that includes noise”) from one of the multiple sources to a set of features of a specified pattern printed on a specified substrate; (Mack, ¶0045: “sources of variation that affect patterning fidelity (e.g., exposure dose and focus variations, hotplate temperature non-uniformity, scanner aberrations”) and executing the machine learning model to determine a classification associated with the specified dataset, (Pandev, ¶0006: “A model is trained that specifies a relationship between the first plurality of electron-microscope images and the defect classes”) wherein the classification identifies a specified source (Pandev, ¶0031: “a first defect class corresponds to a first type of defect that is visible on a first wafer”) of the multiple sources as the source of error contribution for the error contribution values in the specified dataset. (Pandev, ¶0035: “defects on the one or more semiconductor wafers are predicted (216) using the model. Defects on the one or more semiconductor wafers are thus classified”). The proposed combination as well as the motivation for combining Mack, Ramos and Pandev references presented in the rejection of claim 21, apply to claim 24 and are incorporated herein by reference. Thus, the apparatus recited in claim 24 is met by Mack, Ramos and Pandev.
Regarding claim 25, Mack in view of Ramos and in further view of Pandev teaches, The non-transitory computer readable medium of claim 24, wherein receiving the specified dataset includes: decomposing, using a decomposition method, multiple measurement values associated with the set of features (Pandev, ¶0003: “measure line spacing (i.e., pitch), line edges as shown in a critical-dimension scanning-electron-microscope (CD-SEM) image are averaged and the distances between successive average values are determined”) to derive a collection of datasets representative of error contributions from each of the multiple sources, (Pandev, ¶0035: “defects on the one or more semiconductor wafers are predicted (216) using the model. Defects on the one or more semiconductor wafers are thus classified”) wherein the specified dataset is one of the collection of datasets and corresponds to error contribution from one of the multiple sources. (Pandev, ¶0006: “A model is trained that specifies a relationship between the first plurality of electron-microscope images and the defect classes”). The proposed combination as well as the motivation for combining Mack, Ramos and Pandev references presented in the rejection of claim 21, apply to claim 25 and are incorporated herein by reference. Thus, the apparatus recited in claim 25 is met by Mack, Ramos and Pandev.
Regarding claim 26, Mack in view of Ramos and in further view of Pandev teaches, The non-transitory computer readable medium of claim 25, wherein decomposing the measurement values includes: obtaining an image of the specified pattern; (Mack, ¶0005: “Each image of the set includes one or more instances of a feature within a respective pattern structure, and each image includes measured linescan information corresponding to the pattern structure that includes noise”). However, the combination of Mack, Ramos and Pandev does not explicitly teach, obtaining, using the image, the measurement values, wherein the measurement values are obtained for different sensor values; correlating, using the decomposition method, each measurement value of the measurement values to a linear mixture of the error contributions to generate a plurality of linear mixtures of the error contributions; and deriving, from the linear mixtures and using the decomposition method, each of the error contributions.
In an analogous field of endeavor, Middlebrooks teaches, obtaining, using the image, the measurement values, (Middlebrooks, ¶0045: “measuring device may comprise an optical measurement device configured to measure a physical parameter of the substrate”) wherein the measurement values are obtained for different sensor values; (Middlebrooks, ¶0045: “optical measurement device configured to measure a physical parameter”; sensor in interpreted as an optical measurement device) correlating, using the decomposition method, each measurement value of the measurement values to a linear mixture of the error contributions to generate a plurality of linear mixtures of the error contributions; (Middlebrooks, ¶0071: “contributions 850 from independent sources are determined from the results 810 with optionally reduced number of dimensions. One way to determine the contributions is by independent component analysis (ICA)”) and deriving, from the linear mixtures and using the decomposition method, each of the error contributions. (Middlebrooks, ¶0068: “The contributions (S.sub.1, S.sub.2, . . . S.sub.m) can be determined by determining the matrix A”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Mack in view of Ramos and in further view of Pandev using the teachings of Middlebrooks to introduce decomposing a linear mixture of errors. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of identifying specific sources of errors in the inspected image. Therefore, it would have been obvious to combine the analogous arts Mack, Ramos, Pandev and Middlebrooks to obtain the invention in claim 26.
Regarding claim 27, Mack in view of Ramos and in further view of Pandev and still in further view of Middlebrooks teaches, The non-transitory computer readable medium of claim 26, wherein the different sensor values correspond to different threshold levels associated with the image, (Mack, ¶0117: “the impact of threshold value depends on the specific edge detection method used… one threshold value that minimizes the 3σ roughness measured, with other values changing the roughness quite dramatically”) wherein each measurement value corresponds to a delta critical dimension (CD) value (Mack, ¶0051: “Feature-to-feature size variation caused by stochastics (also called local CD uniformity, LCDU) adds to the total budget of CD variation”) of a feature of the set of features at one of the different threshold values, (Mack, ¶0114: “a conventional threshold edge detection method is used, the range of resulting 3σ roughness values is much greater”) wherein the delta CD value indicates a deviation of a CD value of the feature from a mean value of a plurality of CD values of the set of features. (Mack, ¶0070: “feature-to-feature variation described by the standard deviation of the mean linewidths of the features, σ.sub.CDU(L)”).
Regarding claim 28, Mack in view of Ramos and in further view of Pandev and still in further view of Middlebrooks teaches, The non-transitory computer readable medium of claim 27, wherein each threshold value of the different threshold values corresponds to a threshold of a pixel value in the image. (Mack, ¶0116: “The raw (biased) 3σ LWR (nm) as a function of Gaussian filter x- and y-width (in pixels), using conventional threshold edge detection”).
Regarding claim 29, Mack in view of Ramos and in further view of Pandev and still in further view of Middlebrooks teaches, The non-transitory computer readable medium of claim 26, wherein the measurement values correspond to a local critical dimension uniformity value of the feature (Mack, ¶0051: “Feature-to-feature size variation caused by stochastics (also called local CD uniformity, LCDU) adds to the total budget of CD variation, sometimes becoming the dominant source”) at the different sensor values. (Mack, ¶0054: “CD-SEM settings such as magnification, pixel size, number of frames of averaging (equivalent to total electron dose in the SEM), voltage, and current may cause fairly large changes in the biased roughness that is measured”).
Claims 30-35 are rejected under 35 U.S.C. 103 as being unpatentable over Mack (US 2019/0272623 A1) in view of Middlebrooks et al. (WO 2017/102264 A1 - disclosed by Applicant in the IDS submitted on 11/14/2022).
Regarding claim 30, Mack teaches, A non-transitory computer readable medium having instructions (Mack, ¶0204: “A computer-readable medium storing instructions”) that, when executed by a computer, cause the computer to perform operations for (Mack, ¶0204: “instructions that are executable by a processor to cause a computer to execute operations comprising”) determining error contribution data (Mack, ¶0004: “Lithography models that predict stochastic effects such as pattern edge errors or pattern defects”) (Mack, ¶0082: “Pattern structure 800 includes a substrate 810, such as a semiconductor wafer. A feature 815 is disposed atop substrate”) the operations comprising: receiving image data of a set of features of a specified pattern (Mack, ¶0003: “a set of one or more images. Each image of the set includes one or more instances of a feature within a respective pattern structure”) to be printed on a first substrate; inputting the image data to a machine learning model; and executing the machine learning model to determine error contribution data (Mack, ¶0179: “predicting, by the machine learning model, a variability in a final edge position of a printed pattern due to the stochastic effects”). However, Mack does not explicitly teach error contribution data comprising error contributions from multiple sources to the set of features.
In an analogous field of endeavor, Middlebrooks teaches, error contribution data comprising error contributions (Middlebrooks, ¶0063: “results of the measurement are linear combinations (e.g., the sum) of the contribution from the systematic errors”) from multiple sources to the set of features. (Middlebrooks, ¶0032: “determining contributions from different sources in a set of results measured from a lithography process or a substrate”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Mack using the teachings of Middlebrooks to introduce determining image error contributions from multiple sources. A person skilled in the art would be motivated to combine the known elements described above and achieve the predictable result of addressing the identified sources of errors. Therefore, it would have been obvious to combine the analogous arts Mack and Middlebrooks obtain the invention in claim 30.
Regarding claim 31, Mack in view of Middlebrooks teaches, The non-transitory computer readable medium of claim 30, wherein the image data includes a set of images of the set of features, (Mack, ¶0003: “a set of one or more images. Each image of the set includes one or more instances of a feature within a respective pattern structure”) and wherein the error contribution data includes error contribution values corresponding to local critical dimension uniformity (LCDU) values associated with the set of features. (Mack, ¶0070: “This feature-to-feature variation is called the local critical dimension uniformity, LCDU, since it represents CD (critical dimension) variation”).
Regarding claim 32, Mack in view of Middlebrooks teaches, The non-transitory computer readable medium of claim 30, wherein executing the machine learning model to determine the error contribution data includes: training the machine learning model (Mack, ¶0197: “train the machine learning model”) using multiple datasets, (Mack, ¶0150: “Inputs to the machine-learning algorithm are called training data sets”) wherein the datasets include a first dataset having (a) a first image data of one or more features (Mack, ¶0160: “power spectral density (PSD) dataset representing feature geometry information corresponding to the edge detection measurements of the set of images”) of a pattern to be printed on a substrate (Mack, ¶0082: “Pattern structure 800 includes a substrate 810, such as a semiconductor wafer. A feature 815 is disposed atop substrate”) and (b) a first error contribution data with error contributions (Middlebrooks, ¶0063: “results of the measurement are linear combinations (e.g., the sum) of the contribution from the systematic errors”) from multiple sources to the one or more features. (Middlebrooks, ¶0032: “determining contributions from different sources in a set of results measured from a lithography process or a substrate”). The proposed combination as well as the motivation for combining Mack and Middlebrooks references presented in the rejection of claim 32, apply to claim 30 and are incorporated herein by reference. Thus, the apparatus recited in claim 32 is met by Mack and Middlebrooks.
Regarding claim 33, Mack in view of Middlebrooks teaches, The non-transitory computer readable medium of claim 32, wherein the first image data includes a first image of a feature of the one or more features, (Mack, ¶0005: “Each image of the set includes one or more instances of a feature within a respective pattern structure”) and wherein the first error contribution data includes a first set of error contribution values (Mack, ¶0005: “each image includes measured linescan information corresponding to the pattern structure that includes noise”; and ¶0160: “dataset representing feature geometry information corresponding to the edge detection measurements of the set of images”)) corresponding to delta critical dimension values of the first feature. (Mack, ¶0051: “Feature-to-feature size variation caused by stochastics (also called local CD uniformity, LCDU) adds to the total budget of CD variation”).
Regarding claim 34, Mack in view of Middlebrooks teaches, The non-transitory computer readable medium of claim 32, wherein the first image data includes a first set of images of multiple features of the one or more features, and wherein the first error contribution data includes a first set of error contribution values (Mack, ¶0160: “PSD) dataset representing feature geometry information corresponding to the edge detection measurements of the set of images”) corresponding to local critical dimension uniformity values of the features. (Mack, ¶0051: “Feature-to-feature size variation caused by stochastics (also called local CD uniformity, LCDU) adds to the total budget of CD variation”).
Regarding claim 35, Mack in view of Middlebrooks teaches, The non-transitory computer readable medium of claim 32, wherein the error contributions include: an image acquisition tool error contribution that is associated with an image acquisition tool used to acquire the first image data, (Mack, ¶0053: “a critical dimension scanning electron microscope, CD-SEM) are contaminated by measurement noise caused by the measurement tool”) a mask error contribution that is associated with a mask used to print the pattern on the substrate, (Mack, ¶0070: “sources of error (scanner aberrations, mask illumination non-uniformity”) and a resist error contribution that is associated with a resist used to print the pattern, wherein the resist error contribution includes photoresist chemical noise (Mack, ¶0047: “chemical reactions (including those that occur during a post-exposure bake) change the solubility of the resist, enabling patterns to be developed and producing the desired critical dimension”) and a shot noise associated with a source of a lithographic apparatus used to print the pattern. (Mack, ¶0048: “randomness within small volumes… generally referred to as “shot noise”, and is an example of a stochastic variation in lithography”).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MEHRAZUL ISLAM whose telephone number is (571)270-0489. The examiner can normally be reached Monday-Friday: 8am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Saini Amandeep can be reached at (571) 272-3382. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MEHRAZUL ISLAM/Examiner, Art Unit 2662
/AMANDEEP SAINI/Supervisory Patent Examiner, Art Unit 2662
Prosecution Insights